Glycerol derivatives represent useful synthetic building blocks toward a variety of compounds that are of interest for pure and applied chemical investigations as well as for the development of commercially significant manufacturing processes. For example, as part of an applied program within the pharmaceutical arena to define structure-activity relationships associated with the anticancer drug paclitaxel (Erhardt, P. W. Taxane Journal 1997, 3, 36) there is a need for a ready and inexpensive synthesis of 2-phenylglycerol, 1 as shown below where R=.O slashed.. ##STR1##
Specifically, the inventors and others have been trying to define the role that might be played by the 3-acetoxyoxetane system within paclitaxel toward producing the latter's beneficial anticancer properties (Chen, S.-H.; Huang, S.; Wei, J.; Farina, V. Tetrahedron 1993, 49, 2805). This oxetane system is a distinctive molecular arrangement (Macias, F. A.; Molinillo, J. M. G.; Massanet, G. M. Tetrahedron 1993, 49, 2499) that, along with its potential formation within the novel context of 2-phenylglycerol, takes on considerable interest from a chemical manufacturing point of view (Collado, I. G.; Macias, F. A.; Massanet, G. M.; Molinillo, J. M. G.; R.-Luis, F., J. Org. Chem. 1987, 52, 3323 and Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang, F.; Biedeger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L. N.; Liu, J. H. Taxane Anticancer Agents Basic Science and Current Status, 288; Georg, G. I.; Chen, T. T.; Ojima, I.; Vyas, D. M., Eds.; American Chemical Society: Washington, D.C., 1995).
Anticipating a ring closure similar to that developed by Holton et al., 1994 during their total synthesis of paclitaxel, (Holton, R. A.; Somoza, C.; Kim, H. B.; Liang, F.; Biedeger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L. N.; Liu, J. H. J. Am. Chem. Soc. 1994, 116, 1597 and Holton, R. A.; Kim, H. B.; Somoza, C.; Liang, F.; Biedeger, R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L. N.; Liu, J. H. J. Am. Chem. Soc. 1994, 116, 1599) the inventors initially deployed acetophenone, 2, according to FIG. 1 as the most inexpensive and directly analogous starting material appropriate for producing 2-phenylglycerol, 1 where R=.O slashed., which could then be used as the key building block for further elaboration into the desired oxetanes. Toward this end, the enol form of acetophenone, 2, was trapped as its trimethylsilyl (TMS) ether (Step a) (Rubottom, G. M.; Gruber, J. M.; Juve, Jr., H. D.; Charleson, D. A. Org. Synth. 1986, 64, 118) which was vacuum distilled (42-43.degree. C., 0.10 mm Hg) prior to its oxidation with meta-chloroperbenzoic acid (Step b) to produce .alpha.-hydroxyacetophenone. In addition to having appropriate NMR spectroscopic data, crystallization of the latter from hexane/ethyl acetate (1/1) provided material having a melting point of 88-90.degree. C. which is comparable to the technical literature (Yorozu, K.; Takai, T.; Yamada, T.; Mukaiyama, T. Bull Chem. Soc. Jpn. 1994, 67, 2195: melting point=87 to 89.degree. C.) as well as to commercial grade specifications (Aldrich Chem. Co.: melting point=86 to 89.degree. C. and Acros Organics/Fisher Scientific: melting point=86 to 88.degree. C.). Formation of the TMS ether (Step c) followed by immediate reaction with methylmagnesium bromide (Step d) then produced racemic 1,2-dihydroxy-2-phenylpropane. Flash chromatography and crystallization of the latter from hexane/ether (1/1) provided material again having appropriate NMR data and a melting point of 43-45.degree. C. which is comparable to both the technical literature and commercial grade specifications (Eliel, E. L.; Freeman, J. P. J. Am. Chem. Soc. 1952, 74, 923: melting point=44 to 45.degree. C. and Aldrich Chem. Co.: mp=44 to 45.degree. C., respectively). However, at this point the overall yield for just the first four steps of this eight-step strategy had already plummeted to less than 20%. Thus, a higher yielding and more expedient pathway to compound 1 was sought and the latter steps of this first approach, as shown by dotted lines in FIG. 1, were not pursued.
Expediting the synthesis by deploying .alpha.-methylstyrene, 3, as an equally inexpensive and appropriate starting material was tried next according to the scheme depicted in FIG. 2. Starting material 3 was gently oxidized by using sodium perborate (NaBO.sub.3) in acetic acid (Step a) to produce 1,2-dihydroxy-2-phenylpropane whose 1-position hydroxyl group was simultaneously protected via an acetyl function when the procedure was conducted according to Gupton, et al. 1988 (Gupton, J. T.; Duranceau, S. J.; Miller, J. F.; Kosiba, M. L. Synth. Comm. 1988, 18, 937). Purification of the initial material from this reaction was accomplished by vacuum distillation in a range (96-98.degree. C., 0.25 mm Hg) similar to that specified within the technical literature (Gupton et al. 1988) to provide a high quality product in about 40% yield. This is itself a significant improvement since this product corresponds to the same type of intermediate that would have been obtained after five steps (Steps a through e) according to the previous strategy as shown in FIG. 1. Dehydration of the tertiary alcohol (Step b in FIG. 2) with either Burgess reagent (Burgess, E. M.; Penton, Jr., H. R.; Taylor, E. A. J. Org. Chem. 1993, 58, 26) or hexamethylphosphoramide (Monson, R. S.; Priest, D. N. J. Org. Chem. 1971, 36, 3826) proved less effective than by simply heating the material at 165.degree. C. in DMSO for 10 hours (Traynelis, V. J.; Hergentrother, W. L.; Livingston, J. R.; Valicenti, J. A. J. Org. Chem. 1962, 27, 2377). Purification by vacuum distillation (60-66.degree. C., 0.05 mm Hg) was less effective than by flash silica gel column chromatography eluted with toluene to provide a material of acceptable quality (appropriate NMR spectrum with no extraneous peaks) in about a 50% yield. Oxidation was then smoothly effected by again utilizing NaBO.sub.3 in acetic acid (Step c), this time affording a mixture of the mono- and di-acetylated primary alcohol versions of the desired 2-phenylglycerol in about 35% yield. Although these materials were readily separable by flash silica gel column chromatography when eluted with hexane/ethyl acetate (1/1), they tended to reform mixtures even when stored at 2 to 3.degree. C. This propensity for acylated glycerols to readily undergo isomerization by transesterification processes is also exemplified by diacylglycerol which is sold by Aldrich Chemical Co. under the name `Diacetin` as a mixture of the 1,2-, the 1,3- and the 1,2,3- di- and triacetylglycerols. Therefore, upon repetitions of the method shown in FIG. 2, the mixed products from the second NaBO.sub.3 oxidation reaction were directly subjected to hydrolysis using 1 N NaOH in methanol (Step d). Purification could then be much more effectively accomplished by crystallization from CH.sub.2 Cl.sub.2 /CCl.sub.4 (3/1) to provide the final product 2-phenylglycerol, 1 with R=.O slashed., in about a 70% yield for the final step and in about a 5% overall yield for the entire four-step procedure. Complete physical property specifications for 2-phenylglycerol are provided in the Examples Section.
Simultaneous with our investigations, FIG. 3 depicts a very similar four-step process to produce 2-phenylglycerol that was recently reported by Goodall et al. 1997 (Goodall M.; Kelly, P. M.; Parker, D.; Gloe, K.; Stephan, H. J. Chem. Soc. Perkin Trans. 2 1997, 59). However, this strategy deploys the more expensive .alpha.-bromoacetophenone, 4, as the starting material to arrive at the same substituted styrene intermediate similarly obtained after two steps within FIG. 2. Furthermore, OsO.sub.4 is then utilized for the oxidation followed by sodium methoxide for the hydrolysis, both reagents being far less practical from a manufacturing point of view than those deployed in FIG. 2 to effect these same types of conversions. Nevertheless, the reported overall yield in the Goodall et al. case is nearly 30%.
Continuing to explore synthetic strategies that might provide 2-substituted glycerol materials in an even more expeditious and higher yielding manner while still keeping the costs down, we have now discovered that 1,3-dihydroxyacetone can be subjected to an addition reaction scheme, such as that afforded by a simple Grignard reaction, so as to produce 2-substituted glycerols with a broad range of substituent possibilities. Thus, according to the present invention, a novel, two-step strategy, as shown by FIG. 4, can be used to very conveniently obtain 2-phenylglycerol in nearly 70% yield after deploying the inexpensive dimer of 1,3-dihydroxyacetone as the starting material.